Top 10 AI Prompts and Use Cases and in the Government Industry in Boise

AI matters for Boise government because it can speed routine tasks, improve resident-facing services, and surface policy insights - if deployed with clear guardrails and local expertise.

Boise State's CI+D is building campus-wide capacity around generative AI and public workshops to ground those conversations (Boise State CI+D generative AI initiatives), while the City of Boise formalized rules (effective 12‑1‑23) that require IT approval, human validation of generated content, disclosure of AI use, and avoidance of sensitive data to protect residents (City of Boise AI regulation policy (effective Dec 1, 2023)).

Local programs that combine training, ambassadors, and cross‑agency pilots - models highlighted as best practice for spreading AI across city halls - have driven rapid uptake (a reported 10× increase in estimated employee use), showing Boise can gain real productivity while managing privacy and equity risks (Strategies for spreading AI throughout local government (Bloomberg Cities)).

Bootcamp	Length	Early-bird Cost	Register
AI Essentials for Work	15 weeks	$3,582	Enroll in the Nucamp AI Essentials for Work bootcamp

“AI is a technology for which top-down adoption just isn't going to be effective.” - Kyle Patterson

Selection focused on practical, low-risk prompts that map directly to Boise's existing governance and local needs: each prompt had to satisfy the City of Boise's requirements for IT approval, human validation of generated content, disclosure of AI use, and explicit prohibition on sharing sensitive data (see the City of Boise AI regulation for details); vendors' training-data provenance and the ability to disable AI features were treated as pass/fail gating questions.

Prompts were prioritized for public-facing impact (communications, translation, chatbot flows), operational value (meeting minutes, procurement checklists, dataset analysis) and measurable outcomes tied to local metrics, drawing on Nucamp guidance on metrics to measure AI success in Boise.

The result: a compact set of use cases that enforce Boise's four cardinal rules while targeting quick wins - each candidate must state who reviews outputs and how often the system will be audited, a concrete step that keeps human judgment central.

• IT approval & vendor review: Prevents unauthorized tools and requires training-data provenance.
• Human validation required: Ensures accuracy for public communications.
• Disclosure of AI use: Builds trust and aids error detection.
• No sensitive data in prompts: Protects resident privacy and legal compliance.

References: City of Boise artificial intelligence regulation and policy; Nucamp guidance on metrics to measure AI success in Boise.

When using the prompt "Summarize and verify accuracy of this draft public communication; flag legal, privacy, or bias issues," instruct the model to produce a short summary of intent and audience, extract every factual claim with suggested source checks, and output a prioritized list of legal/privacy/bias flags (e.g., potential PII exposure, unclear attribution, copyright risk, or demographic bias).

Follow Boise's rules: every AI-generated or AI-assisted public message must be fact-checked and validated by a person, avoid placing sensitive data in prompts, and disclose AI use and model version where significant communications are involved - see the City of Boise AI regulation for public communications and disclosure requirements.

Include an explicit reviewer field and edit-level, using the regulation's example disclosure:

"This document was generated by ChatGPT 3.5 and edited (heavily | moderately | lightly) by John Doe"

so errors can be traced and IT approval documented; pair this workflow with local success metrics to measure accuracy and trust over time (Nucamp AI Essentials for Work syllabus and metrics for measuring AI success).

Use the prompt

Generate an accessible version of this document (plain language, captioning, alt text) and identify accessibility gaps

to produce a short plain‑language rewrite, machine‑ and human‑checked alt text for images, verbatim captions and a timestamped transcript for video, plus a prioritized gap analysis tied to Boise's Title II obligations and local findings - flag items such as older videos lacking captions, inaccessible fillable web forms, unclear grievance access, and limited staff awareness of auxiliary aids; include explicit remediation owners (e.g., Community Accessibility Manager) and timelines so the city can act (the ADA Transition Plan documents remediation of 2,381 barriers through 2028).

Pair each recommended fix with how to verify it (WCAG 2.0 AA or Section 508 checks), a reviewer field, and a simple public disclosure line describing AI assistance.

For local examples and requirements, reference the City of Boise Self‑Evaluation and the Cross Disability Taskforce guidance, and align plain‑language and translation steps with the Boise Police LEP Assessment to ensure vital documents reach LEP residents.

Accessibility gap (Boise findings)	Recommended action
Older videos lack captions	Provide accurate captions/transcripts; verify before publishing
Website and fillable forms not fully accessible	Train web managers; vet third‑party tools for WCAG 2.0 AA/Section 508
Limited awareness of auxiliary aids and LEP services	Publish request process, train staff, use Language Line and I Speak cards
Large number of physical barriers (2,381)	Prioritize fixes in ADA Transition Plan with owners and deadlines

Analyze this dataset for anomalies, trends, and produce a short report with visualizations and recommended policy actions

to generate an executive summary, time‑series and distribution plots, a prioritized anomaly list with probable causes, and clear next steps that align with Boise's governance: first verify training‑data provenance and check for PII or shared vendor access, then require human validation before any public‑facing decision - per the City of Boise AI regulation (effective Dec 1, 2023).

Visual outputs should include trend lines, control charts for anomaly windows, and simple maps or dashboards for location‑based patterns; every flagged anomaly must include suggested verification queries, a named reviewer role (IT or the Director, Innovation and Performance), and a disclosure line if AI assisted the analysis.

Pair these outputs with measurable success metrics from local guidance - accuracy, false‑positive rate, time‑to‑resolution - and a documented audit cadence tied to the regulation's review cycle and vendor disable/rollback checks (Nucamp metrics to measure AI success in Boise) so anomalies become governance triggers, not surprises.

Required check	Recommended action
Training-data provenance	Document source, public vs. city-only access
PII exposure	Redact or remove before prompts; route to IT if needed
Disable/rollback options	Confirm vendor can disable AI features; test fallback
Human validation & disclosure	Assign reviewer, record edit level, publish disclosure
Audit & review	Schedule periodic audits and annual policy review

Design citizen‑facing chatbot flows so every conversation follows Boise's AI rules: open with a plain‑language greeting and intent menu, collect only non‑sensitive context (never prompt for SSNs or other PII), run clarifying questions, and surface a clear “Talk to a human” fallback that supplies department contact details and an explicit reviewer field and disclosure of AI use per the City of Boise AI Regulation (Dec 1, 2023).

Build accessibility and escalation into the flow: offer auxiliary‑aid options (qualified sign language interpreters, Braille, TTY), an easy path to request those services, and an escalation path that includes the Community Accessibility Manager's contact so staff can intervene when needed, following Boise's Boise ADA and Section 504 Notice.

For fallbacks, prepare canned responses that explain limits of the bot, a named human reviewer to verify any AI‑generated guidance, and a vendor/IT checkpoint before deploying new features - this keeps resident trust and compliance front and center.

Escalation contact	Phone / TTY / Email
Community Accessibility Manager	(208) 972-8573 • TTY: (800) 377-3529 • communityengagement@cityofboise.org
Human Resources / AI oversight	(208) 972-8090 • TTY: (800) 377-3529

Turn a long Boise meeting transcript into a compliance-ready set of minutes by first producing a clean, speaker-identified transcript (use real-time tools that support Zoom/Meet/Teams) and then pasting it into an AI prompt such as: “Write formal meeting minutes from the transcript below - include meeting title, date/time, attendee list, 3-bullet executive summary, key decisions, and action items in the format ‘Assignee: Task - Due date' with a named reviewer and disclosure of AI use.” Tools like Tactiq real-time transcription and AI summaries or templates from Otter meeting notes template with action items speed transcription and show the fields to capture; follow local rules by redacting PII, naming the human validator, and adding explicit deadlines so minutes become governance artifacts, not loose notes.

Final step: route the draft to the assigned reviewer (IT or department lead) for validation and publish with the required AI disclosure so residents and auditors can trace edits and decisions.

Required minutes field	Why it matters
Meeting title, date & time	Context for records and audits
Attendees	Accountability and follow-up
Decisions	Official record of outcomes
Action items (Assignee + due date)	Drives completion and metrics
Named reviewer & AI disclosure	Compliance and traceability

“Flippin' fantastic. Best meeting companion I've ever used. Nothing else comes even close.” - Steve Coppola

Make procurement decisions defensible by turning Boise's regulation questions into non‑negotiable contract clauses: require IT approval before purchase and route high‑value buys through Finance‑Purchasing (centralized services for purchases $50,000+); demand written training‑data provenance (public vs.

city‑only), names of parties who create and maintain datasets, and whether vendor or third parties can access city inputs or prompt/response data; insist on clear disable/rollback options and a pre‑deployment test showing the functional impact of disabling AI components; contractually bind vendors to support human validation workflows, annual audits, and timely disclosures whenever AI significantly shapes public communications; and use Purchasing's vendor registration and bid process to collect these responses up front so proposals can be scored on governance as well as cost.

Embed these checklist items in RFPs and require evidence (policy documents, flow diagrams, and IT sign‑off) as pass/fail gates to keep Boise's four cardinal rules enforceable in procurement (see the City of Boise AI Regulation (Section 430Q)) and align contracting steps with centralized purchasing oversight (City of Boise Purchasing Department), turning vendor diligence into a repeatable, auditable part of each acquisition.

Checklist item	Action / Evidence required
IT approval	Signed IT approval before award
Training‑data provenance	Vendor disclosure: sources, public vs. city-only, dataset owners
Data access	Written statement of who can access city data and prompt/response retention
Disable/rollback	Contractual right to disable AI features; pre‑deployment impact test
Human validation & disclosure	Workflow describing reviewer roles and mandatory public disclosure phrasing
Audit cadence	Annual audit schedule and remediation commitments

"This document was generated by ChatGPT 3.5 and edited (heavily | moderately | lightly) by John Doe"

Assessing ethical risks for a proposed Boise AI system starts with three local guardrails: require IT approval and vendor provenance checks, forbid sensitive data in prompts, and mandate human validation plus public disclosure of AI use - rules already codified in the City of Boise AI regulation 4.30q (employee policy) and intended to prevent accidental leakage of resident data into third‑party models (City of Boise AI regulation 4.30q (employee policy)).

Practical mitigation steps map directly to common Idaho concerns from state surveys and guidance: redact or token‑ize PII before any prompt, require vendor contracts to include disable/rollback clauses and dataset provenance, run pre‑deployment bias tests and impact assessments, and assign a named reviewer for every public output.

Audit metrics should be concrete and repeatable - e.g., percentage of prompts redacted for PII, false‑positive/false‑negative rates on bias tests, number of vendor‑reported data accesses, and completion of annual audits and inventory updates as urged by Idaho's state IT framework and legislative working group - so ethical checks become measurable governance actions, not ad hoc promises (Idaho Capital Sun survey on AI use in government agencies, Idaho OITS AI Framework for Artificial Intelligence).

Ethical risk	Mitigation step	Audit metric
Privacy / PII leakage	Redact/tokenize data; ban PII in prompts	% prompts redacted; vendor access logs
Bias / fairness	Pre‑deployment bias tests; impact assessments	false‑positive/negative rates; bias test coverage
Transparency & accountability	Named human reviewer; disclosure of AI use/model	% outputs with reviewer+disclosure
Vendor risk / rollback	Require provenance + disable/rollback clauses	Contracted rollback tests; time‑to‑disable

"If we allow people to put in any sort of protected data into the AI, it becomes part of the large language model. And I think that's the scariest part."

Use the prompt “Translate this city document into Spanish and check cultural appropriateness and readability levels” to produce a human‑reviewable Spanish version, a short plain‑language summary, a list of cultural or dialectal flags (e.g., regional idioms, tone for immigrant audiences, or legal phrasing that may confuse LEP readers), and a readability check tied to actionability - mark any “vital” content (permits, police notices, utility bills) that must be escalated for certified translation and Title VI review.

Require the model to output: (1) sentence‑by‑sentence source alignment for easy verification, (2) suggested alt text and captioning options to meet ADA effective‑communication needs, and (3) a named bilingual reviewer with instructions to confirm accuracy before publication; route interpretation requests through the City's online form and Language Line when needed.

This workflow builds on Boise's current Language Access Resources and the Police Department's LEP plan that already translate many vital documents into Spanish, while recognizing the city's FY24 bilingual pay program that incentivizes qualified staff - so translated materials actually get stewarded, not just produced (City of Boise Language Access Resources, Police Department LEP Assessment and Plan).

Contact / Program	Details
Language Access Program Manager	Danny Galvez - (208) 972-8498 • dgalvezsuarez@cityofboise.org
Bilingual Pay (FY24)	Oral proficiency: $1,500/yr; Oral+written: $3,000/yr - Spanish only for FY24

“We're doing compliance really well. We're doing a good job with the ‘I Speak' [language identification] cards and translating our documents, but we need to go to the next level and make sure we're providing information in the languages and ways of communicating our community is asking for.”

Create a concise, hands‑on staff training package that pairs Boise State's short faculty workshops and microlearning with clear city guardrails: open with the City of Boise AI regulation and policy (IT approval, human validation, disclosure, and no sensitive data in prompts - see the City of Boise AI regulation and policy), then run a 1–3 hour practical session from Boise State's CI+D playbook that includes prompt‑hygiene drills (redact or tokenize PII before any prompt), an assignment‑testing exercise to check whether an LLM can finish student work, and a short checklist for vendor/data provenance and disable/rollback clauses.

Offer an optional deeper cohort via the three‑day AI Institute model to build course‑level integration and ethics review practices (2025 AI Institute for Teaching & Learning participation details), and embed bite‑size followups (microlearning, brownbags) from Boise State's faculty/staff resources so skills stick (Boise State CI+D generative AI initiatives and resources).

Make "who reviews outputs" a mandatory field on every AI task, include a canned disclosure line for public documents, and measure success with simple metrics: % of prompts redacted, time to human review, and number of vendor provenance checks completed - turning training into audit‑ready governance that protects residents while unlocking practical productivity for Idaho government staff.

Training option	Length	Notes / Stipend
CI+D short faculty workshop	1–3 hours	Practical prompts; $50 stipend noted for some workshops
AI Institute for Teaching & Learning	May 19–21, 2025 (3 days)	Hands‑on course design; $250 stipend for selected participants
Microlearning & brownbags	Minutes to 1 hour	Ongoing reinforcement for staff

“AI is like fire - can be a powerful tool or destructive; aim to harness as a tool.”

Simulate AI error scenarios by running tabletop exercises that inject realistic failures - misinformation in a draft press release, dataset drift that flips eligibility decisions, or a mistranslation of a vital notice into Spanish - to produce a simple risk matrix (impact × likelihood), an internal and public communication plan, and clear human‑validation checkpoints before any decision goes live.

For each scenario, map who detects the error, who validates the fix, and how to prove rollback worked (vendor disable/rollback tests are a must); pair those steps with measurable success metrics so detection and recovery become audit trails rather than anecdotes (see Nucamp AI Essentials for Work syllabus on measuring AI success: Nucamp AI Essentials for Work syllabus - metrics to measure AI success).

Include a communications template that names the reviewer, discloses AI use, and directs LEP or disability escalations to bilingual staff or accessibility managers - these simulations should inform procurement and policy choices shaped by Idaho AI policy signals (see Nucamp AI Essentials for Work syllabus: Nucamp AI Essentials for Work - Idaho AI policy guidance) and test risks highlighted by generative translation tools for multilingual services (see Nucamp AI Essentials for Work syllabus on multilingual AI risk testing: Nucamp AI Essentials for Work - testing generative translation risks).

The so‑what: practiced simulations expose gaps - staff, vendor, or disclosure - that can be fixed before a real error affects residents.

Getting started safely with AI in Boise means pairing the City's clear guardrails - IT approval before any tool, mandatory human validation of generated content, explicit disclosure of AI use, and a ban on sharing sensitive data (City of Boise Regulation 4.30q, effective 12‑1‑23) - with a short, practical rollout that builds staff capability and vendor controls; a concrete first step is to stand up a cross‑agency pilot that enforces those four cardinal rules, trains reviewers on prompt hygiene, and measures success with simple metrics (percent prompts redacted, time to review, vendor access logs).

For training and repeatable skills, enroll a cohort in a practical program such as Nucamp's AI Essentials for Work (15 weeks) to standardize prompt hygiene, vendor questions, and human‑in‑the‑loop workflows across departments (syllabus and registration below).

These paired actions convert Boise's policy into day‑to‑day practice so Idaho residents get faster, more equitable services without sacrificing transparency or privacy.

Program	Length	Early‑bird Cost	Register / Syllabus
AI Essentials for Work	15 weeks	$3,582	AI Essentials for Work registration (Nucamp) • AI Essentials for Work syllabus (Nucamp)

“AI is a technology for which top-down adoption just isn't going to be effective.” - Kyle Patterson